Team roles

Analytics teams come in all sizes and shapes. What doesn't change is the need for clarity on who does what. This part outlines the main roles you'll see in a modern analytics function and how they fit together.

Core roles

Analytics lead / Head of analytics — Owns the link between analytics and the business: prioritisation, roadmap, stakeholder management, and team design. Ensures the team is working on the right questions and that outputs land with decision-makers.

Analyst (generalist) — Runs analyses, builds reports and dashboards, and answers ad-hoc questions. Often the primary interface with business users. In an agentic setup, they spend more time on interpretation, quality checks, and high-value work; agents handle a growing share of repetitive queries and reporting.

Data engineer — Builds and maintains pipelines, warehouses, and data models. Ensures data is available, reliable, and documented. Works closely with analysts and (where relevant) ML engineers so that the right data is in the right place.

Data scientist — Focuses on advanced analytics: forecasting, segmentation, optimisation, and bespoke models. In agentic environments, they design and validate what agents do and step in when human judgement or custom work is required.

ML Engineer — Takes models from development to production: builds serving infrastructure, monitors performance, and maintains MLOps (versioning, retraining, A/B tests). Works with data scientists (who design and validate models) and data engineers (who provide pipelines and features). In agentic analytics, they often own the systems that run and scale the models behind agents — reliability, latency, and cost.

Analytics product / platform owner — In larger teams, someone may own the tooling, standards, and ways of working (e.g. self-serve analytics, agentic platforms). Ensures the function scales and stays consistent.

How they work together

Analytics works best when roles are defined by outcomes and handoffs, not by turf. The lead aligns the team to business decisions; analysts and data scientists produce the analysis; data engineering provides the foundation. With agentic analytics, the balance shifts: analysts and scientists spend more time on design, validation, and exception-handling, while agents handle volume and repetition.

Sizing and structure

Small teams often combine roles (e.g. analyst–engineer, or lead–analyst). As you grow, specialisation increases. The goal is to have clear ownership for decisions, data, and delivery — and to avoid gaps where "everyone thinks someone else does it."